Q$1.$ A data mining routine has been applied to a transaction dataset and has classified $88$ records

as fraudulent $(30$ correctly so$)$ and $952$ as non$-$fraudulent $(920$ correctly so$).$

a$)$ Manually construct the confusion Matrix and calculate the overall accuracy rate.

b$)$ Another analyst comments that you could improve the accuracy of the model by

classifying everything as nonfraudulent. If you do that, what is the accuracy rate? Create

the corresponding confusion matrix. Does the error increase or decrease compared to

that in part a$?$

c$)$ If the manager comments that the fraudulent transactions are very costly to the

company and he$/$she asks you to compare the model in a$)$ with the model in b$),$ which

would you recommend and why?